| dc.description.abstract | The market for biometric technology continues to grow. Biomet rics is used in the real world in a wide variety of fields, from surveillance, health care, advertising, Human-Robot Interaction to security and trust. Interest in this field is transversal. Soft biometrics have emerged in recent years as a potential alternative to and useful ally of primary biometrics (also known as hard biometrics). Any anatomical or behavioral feature that gives some information about a person’s identity but does not provide sufficient evidence to accurately determine identity can be called a "soft biometric trait." In addition to the evidence that these characteristics can be used to improve the accuracy of a recognition system, the study of soft biometrics has shown that additional information about people, such as age, gender, eth nicity, and information about emotional and cognitive state, can be inferred from these soft data, demonstrating their broader potential. Based on this, it is a study area that has garnered considerable interest over the past decade but has never quite taken off. In fact, there has been an absence of a more verticalized and system atic examination of certain characteristics and their purposeful and widespread use in applications such as HRI. For this rea son, after a thorough examination of the literature, this thesis focused on two aspects. First, the potential of soft biometrics when integrated with the native capabilities of social humanoid robots was studied to demonstrate how their application can make this type of robot even more effective in the sectors of elderly care, security, and also become a danger in the area of So cial Engineering. Then, periocular soft biometric features (blink, eye-movements, fixations, and pupil) were studied in detail to demonstrate their potential for the purposes of recognition, de mographic classification, emotion detection, and cognitive state analysis. [edited by the Author] | it_IT |
